A quantitative evaluation of an anti-cancer vaccine for treating advanced prostate cancer
In this talk, I will present our mechanistic model of the response of prostate cancer to one of the the first FDA-approved live cell anti-cancer vaccines, sipuleucel-t (Provenge). In clinical trials, Provenge has shown only modest survival benefits. Moreover, an optimal dosing schedule has not been established, even after a decade of use. Our model is calibrated with data from mouse xenograft experiments, and captures the detailed immune response of the body when vaccination is administered. I will then introduce our modeling paradigm — Standing Variations Modeling — which captures the inherent heterogeneity that characterizes individuals in a population, and provides an explanation for the observed clinical outcomes of treatment with Provenge. We also predict an optimal therapeutic regime that maximizes predicted efficacy of the vaccine for a small subset of a heterogeneous population. Our approach readily generalizes to a range of emerging cancer immunotherapies, and more generally, to predicting and understanding how a population responds to any intervention targeting a human disease. An alternative approach to capturing heterogeneity in a population is using agent-based models (ABMs), where each cancer cell is an independent agent. Time permitting, in the second half of the talk, I will present some current research directions wherein we have developed a novel method to parameterize computationally complex ABMs of tumor-immune interactions with coarse-grained and noisy experimental data.